Abstract
Neely, Walter, Black, and Reiss (2012) have done a service in identifying brain regions involved in humor appreciation in children's developing brains. The point of this commentary is to address some problems that should be acknowledged and improved by future researchers in this field.
Humor, a cognitive-affective style which includes the capacity to perceive, interpret, enjoy, create, and relay incongruous communication (Samson, Zysset, & Huber, 2008), is described as one of the transcendent character virtues that forges connections to the environment and provides meaning (Peterson & Seligman, 2004). Humor should be acknowledged as a field of study, not a mere topic, since it has many aspects requiring expert consideration. This complexity is reflected in the multitude of different theories of humor.
This commentary will focus on explaining humor from the perspective of previous neuroimaging studies. The aspects of “humor” focused on in these studies are the cognitive and emotional processes taking place within an individual. This cognitive model is an attempt to explain the necessary conditions and neural dynamics for eliciting a reliable humor response. The goal was to identify a universal cognitive process.
The first step posited in the cognitive humor model is situation encoding. The situation is defined as the initial information and all of the semantic attachments connected to that information by the perceiver (Vaid, Hull, Heredia, Gerkens, & Martinez, 2003). The situation can be unambiguously provided (e.g., a joke is directly told) or be perceived and interpreted spontaneously in the environment. Following this initial encoding, the second step is to detect an incongruity and the third is to resolve that incongruity. Incongruity is defined as the presence of at least two potential meanings that are incompatible (Suls, 1972), i.e., when information is presented that is inconsistent with the initially encoded situation (e.g., punchlines and benign grammar violations). Resolution is the integration of the incongruous elements (Suls, 1972), made possible by the individual's knowledge about the attributes of the incongruous elements. In resolution, an understanding of why the situation is humorous is clarified. The final step of humor processing posited by the cognitive humor model is humor appreciation. If the humor is enjoyed, positive emotion is elicited (defined as “mirth” in the literature).
Current neuroimaging research has been teasing out which brain areas may be involved in each of the above processes. The most-replicated findings are related to the involvement of the left inferior frontal gyrus (left IFG) and bilateral temporo-occipito-parietal junction (TOPJ). The left IFG has been implicated in every humor neuroimaging study examining brain activity during presentation of “funny” stimuli as compared to specific baselines or controls. Regardless of the medium of the joke (i.e., video, drawn, or written), the left IFG appears to be activated. This area has also been identified in the literature dealing with ambiguous semantic knowledge and detecting double meaning in sentences (Bekinschtein, Davis, Rodd, & Owen, 2011) and so seems to be involved with incongruity detection as well as resolution of humor. The bilateral TOPJ is an area where three different lobes meet (including Brodmann areas 39, 40, and posterior parts of 22) and whose cytoarchitecture is greatly heterogeneous. How the TOPJ is involved in processing of humor has been a point of contention: some research groups argue that it processes only incongruity detection (Mobbs, Greicius, Azim, Menon, & Reiss, 2003; Samson, Hempelmann, Huber, & Zysset, 2009) and others that it processes only humor resolution (Azim, Mobbs, Jo, Menon, & Reiss, 2005; Wild, Rodden, Rapp, Erb, Grodd, & Ruch, 2006). Other general functions may include bringing stored expectations online (Moran, Wig, Adams, Janata, & Kelley, 2004), processing general social information (Watson, Matthews, & Allman, 2006), and as a hub where integration of incongruity occurs (Kohn Kellermann, Gur, Schneider, & Habel, 2011).
An important aspect of neuroimaging research in humor is that the related neural activation has only been studied in adults. In terms of developmental theory of humor, research has shown that children enjoy humor that challenges their developmental stages of cognition and social competence (Zigler, Levine, & Gould, 1966). Humor stimuli that are either too simple or too complex for the child's cognitive developmental stage will not be perceived as funny (Pinderhughes & Zigler, 1985; Johnson & Mervis, 1997). How these processes occur alongside brain development has not been investigated until the pioneering work of Neely, Walter, Black, and Reiss (2012).
Neely, et al. (2012) recorded functional MRI (fMRI) while children (N = 15; age range 7.9 to 11.7 years) watched a number of 4- to 16-second videos. These videos were categorized as funny, positive, or neutral based on previous pilot testing in age-matched samples. After the fMRI session, the children rated the videos on scales of funniness and enjoyment. Three major comparisons were made of brain activity in the funny, neutral, and positive conditions.
Relative brain activity was compared during the funny compared to the neutral videos. Only the bilateral TOPJ was significantly more activated. Areas of mesolimbic regions were marginally more active but this difference was not significant, controlling for multiple comparisons. The authors concluded that the TOPJ and mesolimbic areas are part of a humor-processing network already established in childhood and that the TOPJ is involved with resolution and the mesolimbic activity is involved with humor appreciation.
Comparison of brain activity between funny and positive videos had not been contemplated in previous humor neuroimaging work. Bilateral TOPJ activity was greater in response to the funny videos, leading the authors to conclude that this brain area may be specifically used for humor processing, not general reward processing. The third comparison was an attempt to clarify the development of brain activity in processing of funny and positive videos. Although the exploratory analysis did not correct for multiple comparisons (and therefore is in danger of capitalizing on chance), the analysis showed comparatively greater activation in right IFG and mesolimbic areas in younger participants.
Interpretive Issues
Repeated references to reward processing in Neely, et al. (2012) indicate interpretive problems. The two most important inferences related to reward were related to (1) how the identified activity in mesolimbic areas implies reward processing and (2) whether the positive stimuli were rewarding stimuli that were not funny. A short introduction on reward processing will be useful to provide context.
Broadly speaking, rewards are stimuli that increase the probability of behavior (Berridge & Kringelbach, 2008). While things like food, water, and sex are examples of primary rewards, almost any kind of stimulus can acquire reward value through conditioning (Berridge & Kringelbach, 2008). The most common misunderstanding about the reward system is when people equate reward processing with pleasure processing. Reward processing is a much more complex system that involves both conscious and unconscious pleasure perception, motivation, and learned associations towards a stimulus as well as predictions to increase future exposure to the stimulus (Berridge & Kringelbach, 2008).
Certain brain areas have been shown to be involved with different aspects of reward processing. To be brief, only the “mesolimbic areas” will be discussed. While Neely, et al. (2012) never operationally defined what they meant by “mesolimbic reward areas,” based on their previous work (Mobbs, et al., 2003) they likely were referring to areas in the ventral striatum including the nucleus accumbens (NAcc). They likened the ventral striatum with reward processing or “pleasure” associated with humor appreciation. This is important because later these authors associated the ventral striatum to processing unexpected rewards, as would occur in incongruity detection. This is the understanding in most of the literature related to function of the ventral striatum (Schultz, Tremblay, & Hollerman 2000). Neely, et al. seem to conflate two different components of the model as “eliciting reward processes,” an oversimplification of this complex process. A great deal of effort would be required to identify how each part of the reward system is involved with particular aspects of the cognitive model of humor.
The positive category of videos was not specified clearly enough to support their inferences. The videos in this category were included, based on pilot ratings high on “enjoyment” and low on “funniness.” These videos are called the “rewarding” videos throughout Neely, et al. (2012); the assumption seems to be that positive videos have reward value equal to that of the funny videos but have significantly less “funniness.” Neely, et al. designed the funny versus positive comparison to distinguish brain activity specific to processing humor or “funniness.” The term “positive” requires careful definition since there are a multitude of different positive emotions. Because the examples of positive stimuli included snowboarding, a person jumping over cars, and breakdancing, certain candidates' positive emotions for these stimuli could be described as excitement or elevation (awe-inspired feeling). How this emotion compares to mirth or any part of the humor system is not theoretically understood; but in any case, Neely, et al. have provided the first comparison of brain activation for these processes.
Another difficulty comes from the way that the stimuli were analyzed. Neely, et al. (2012) is the third study that used dynamic stimuli (as opposed to static images or written jokes) as humor content (Moran, et al., 2004; Franklin & Adams, 2011). The other studies used longer videos (e.g., 30-minute TV episodes and stand-up comedy specials) while Neely, et al. used short videos between 4 and 16 seconds in length. A very important difference between Neely, et al. and the other studies is that the two previous studies used methods to time-lock the neural activity of interest around the funniest parts of the clip. What time-locking attempts to do is to segment the parts of the stimulus similarly to the aspects of the Cognitive Humor Model. One way to do this is to model the stimulus around the punchline. Without this time-locking, it is not possible to make inferences about how certain brain areas' activation are responses to specific parts of the stimulus and therefore related to specific aspects of humor processing. In fact, Neely, et al.'s explanation for not finding significant ventral striatum activity was that the funny events in each video occurred at different time points and the videos were different lengths. Why they did not time-lock the fMRI with the funny moments of videos was never explained. These problems necessitate cautious interpretation of the comparisons between “funny” and other conditions; it is likely that the activity signal is buried under the variance due to lack of time-locking.
There are additional minor issues in the context of theories and models of humor. Neely, et al. (2012) is the first neuroimaging study on humor not to detect greater activity in the left IFG in any of their comparisons. One possibility for this is the lack of time-locking. Another possibility for the lack of left IFG activity, is that it is not until age 10–11 that children begin to focus on language in processing humor (Semrud-Clikeman & Glass, 2010). Neely, et al. do not discuss not finding left IFG activity. Further humor research should attempt to address the role of left IFG and whether it is involved in humor processing in adolescents.
Problems in Neely, et al.'s (2012) design weaken the developmental interpretation. The first issue is that while the age range was stated (age range = 7.9–11.7 yr.), the distribution of the 15 participants within that range was not stated. A skewed age distribution could have affected the regression analysis in the funny versus positive comparison, e.g., reducing the statistical power. Another issue is that the older the participants were, the less funny they found the funny stimuli. Neely, et al. acknowledged this could be due to differences in brain development or the age-appropriateness of the stimuli, but these two possibilities are confounded. A study using age-appropriate humor stimuli in a sample of adolescents could assess how cognitive mastery is related to humor development.
Neely, et al. (2012) deserves the attention of both developmental and humor experts. This work helped fill a gap in the literature and provides preliminary data from which future studies can be developed. The choice of age range in this study (6 to 12 years) is particularly interesting because it appears to be during this period in development where there is an important transition between enjoying incongruity without resolution to enjoying incongruity with resolution (Semrud-Clikeman & Glass, 2010). Regardless of age group, past research has shown that children enjoy humor that challenges their developmental stage (i.e., the cognitive mastery hypothesis). Different age groups' responses to various humor stimuli should be a focus of future research. Utilization of a detailed model of reward processing would correct some of the confounds from Neely, et al. Lastly, it is hoped that research can continue to characterize, differentiate, and explore the humor processing throughout the lifetime.